autofocus vs manual focus

[reaction]

This is not a question about when. This is a question about how.

AF works by an AF sensor that detects where focus should be. What happens after this? Does it tell the lens/camera screw how far to turn, and assume it’s OK? Or does it say move this direction, and keep checking/moving until it finds good focus?

I’m trying to understand why a lens that’s fine with MF can be off in AF. Does accuracy depend on whether the lens is screw focus, motor focus, or USM focus? :confused013

Also, how do you MF quickly? On a crop body, it takes me 1-2s to find the sharpest point. This makes me useless trying to MF people at an event.:Doh:

[bigdazzler]

Cant actually help with the mechanics mate (I have very little clue on how these cameras and lenses actually work .. I just take pictures :D) .. but I can tell you that I would never ever try and MF on a moving subject, like a person at an event. I would imagine that would be close to impossible. Id be interested to hear if anyone does actually work this way though and how they achieve it.

[FallingHorse]

I can't be of much help either, I rarely use MF too - I have found the only time I ever use it is at night time when the AF tends to go on a 'hunting expedition'. I haven't even attempted to use MF on moving subjects as I am sure the action would be over by the time I found it! I know there ae people that can work in MF but sadly, I'm not one of them!!

[Steve Axford]

Cant actually help with the mechanics mate (I have very little clue on how these cameras and lenses actually work .. I just take pictures :D) .. but I can tell you that I would never ever try and MF on a moving subject, like a person at an event. I would imagine that would be close to impossible. Id be interested to hear if anyone does actually work this way though and how they achieve it.
I believe that photographers at the Grand Prix (for example) will focus on a point and wait for the car to reach that point. They will then fire off as many shots as they can in the hope of getting one in focus. Since a lens cannot track an object that is moving close to 300km/hr, they have little choice. I remember a shot at a Melbourne Grand Prix with a car flying upside down above another car on the track. Good shot, but this was done this way. Take 1000 shots like this and you may get a good one.

P.S. Autofocus requires an interaction between the camera and the lens. The camera decides what is in focus and tells the lens what to do. Cameras can make mistakes.

[bigdazzler]

Yea good example Steve, I see how that would be the case for those guys. I would imagine though at an event, such as a wedding, or similar, where people are moving around unpredictably it would be quite difficult to anticipate focus.

[TOM]

yep, zone focusing. i use it all the time at events, particularly weddings. ie, bride walking down the isle. i will pick a point on the floor, focus to it, and when the bride gets to that point, BAM. if the shot isn't great, then i'll abstain, then move with the subject until the shot is right.

it is a bit harder to focus manually with a modern dslr, as they are geared towards af, and the focus lock has some inaccuracy. but remember, before af, there were millions of photos taken of moving subjects. it is not impossible, in fact far from it. when you get yourself out of the af mindset, you will see that it can be easier to mf in a lot of situations. usually with small format, you could set your diaphragm to f8 on a 35mm lens, and you've got so much DOF. here you would get to know you lens focus ring (hopefully you have a tab on it), so that when you turn it, you know at what distance it is focused to.

it all sounds hard, but you have full control, and with a bit of practice, you'll make all of these decision in the blink of an eye.

[bigdazzler]

But again TOM your example is a situation where you KNOW the bride will be at that point at a certain time, making it easier to anticipate, and as you say, choose a focal point. What about candid stuff, like at receptions ?? Is there a technique or would you revert back to AF for that ?

[arthurking83]

for fast manual focus, you'll need a 'proper' focusing screen.

any of the aftermarket focusing screens available will be better than the standard focusing mattes in a consumer grade DSLR.

I dunno about pro level DSLR bodies, but the focusing screen on DSLRs is optimised for an aperture value of f/5.6, so if you set the lenses aperture to less than this, you're being deceived by the DOF that you can see through the VF.
A split prism or micro prism collar can allow you to see the exact point of focus.
Some aftermarket focusing matte screens are optimised for much faster aperture values(eg. Katzeye with the Plus option).

As I understand it(but never read any tech specs on the topic), the focusing sensor tries to determine the best contrast for a given area, descibed by the line(s) within the AF area box. The box itself is technically useless, as the box contains either one or more lines(depending on whether it's cross type or not). In that AF area box are one or more lines that the AFsensor uses to determine the best contrast(difference between various colours, etc.. kind of like when you USM a part of an image and it becomes sharper, or more defined).
If there are multiple contrasting parts to a scene in that box, eg as you get in a tree branch with a chaotic twig arrangement, it's easy to fool the AF, as it will try to focus on one part and then another, etc, etc and it will focus hunt in many of these situations.
The way I see it, is that the lines(crosshairs) in the AF area act in a similar manner to a split prism, because.
Note, this is how I've understood the system to work: if you point an AF area box with a single vertical line, that line will determine contrast much quicker and easier if the contrast in perpendicular to the AF sensor line. It will still detect contrast if the contrast is parallel to the sensor line, but it can struggle in those situations. Not really a test you can easily produce because you don't know exactly where the sensor line is within the AF area box(it's supposed to be dead centre, but that's not guaranteed).

if you read some AF tech specs in camera body features, you may notice differences in how many 'crosshair' type sensors the camera may have. I think The Canon 1Dmk4 ahs the most of any camera currently available, even tho the current Nikon D3/Dxxx series cameras I think have more sensors.
(or maybe that's the other way around?? :o)
Crosshairs sensors have sensor lines that look like + symbols, single line sensors look like | straight lines.
(I think the Sony A900 has a central sensor that is crosshatched #(like that) which should give faster more accurate contrast detection in AF).

if you want more accurate MF, shoot at f/5.6, get an aftermarket focusing screen, easily user replaceable, and I have some images ready to go that I want to use in a how to one day ... memory(my ability to remember to do this!) and lack of time prevent me from doing this :p

There's a ton of literature available on how it all works.

There are cheap thirdparty focusing screens available on ebay, but I'd recommend against them. I have one for my D70s(which needs it desperately) but it never worked perfectly with the 500/8(split prism blackout!), so I removed it after some time, and now my son uses the camera.
I'm going to get myself either a Katzeye Plus, or one form Focusing Screen.com for my D300(if I can ever get my severely depleted cerebral memory modules to work! :p)

The Focusing Screen.com mattes are half the price and much more configurable compared to the KatzEye screens, but the Katzeye screens work very well(user recommendations) with super fast f1./4, or faster lenses.
Note that much more user configurable means that you determine how you want it set up and then order it. Once you've decided how you want it set up, you're stuck with that screen.
I have no info as to whether the Focusing Screen.com screens black out at f/8, for use with my 500/8 which is one of two lenses that will most benefit from the new screen! That's why I really haven't ordered one yet. I think the more expensive KatZeye may be better in the long run for me.

Hope that helps a little.

[bigdazzler]

and thats why I AF ... all too hard for me AK :eek::D

[arthurking83]

See!! you did really need that A900 after all!

According to a few review sites, it has the best viewfinder of all current DSLRs!

[TOM]

Dazz, I don't use AF at all for 95% of my work. But then again, I don't use SFDSLR. When I do use a SFDSLR, then I do autofocus, just because it is too hard to MF with the lenses that i use and those type of cameras. MF is much quicker and more accurate with a rangefinder, so that's what i use for all of my wedding work. i only zone focus in certain situations, and typically when i'm at f1.4 or thereabouts. i think that autofocus cameras get you into the mode of pin point focusing, where in reality, zone focusing is all you ever need. the example i described above for the wedding is more of a trap focus technique.

AK, i recently tried the d3, 1dsMK4, and the a900 side by side, and the sony did have a better VF. it wasn't something that we were looking for, but it was that apparant that we both commented on it.

[bigdazzler]

See!! you did really need that A900 after all!

According to a few review sites, it has the best viewfinder of all current DSLRs!


actually I got the A850 .. it has a 98% viewfinder as opposed to the A900s 100% , and only shoots a lousy 3.5fps as opposed to the A900s 5fps. They are literally the only differences in the two cameras, and the A850 RRPs for a thousand bucks less !!! Im happy with my decision, and AF works pretty well for me. Im flat out getting decent pictures any time without confusing myself with MF and split focusing screens .. :eek: :lol:

[swifty]

Slightly OT but is what u see in the VF exactly what the sensor plane sees.
I swear there's been occasions where I've very carefully MF but the point of focus is off on the image. This was done on a large medium format viewfinder too but being film I cannot verify if it's my eye/focussing that's the problem or inaacuracies in camera. We're talking very very shallow dof here though.
I guess the gold standard now is mf in the cameras with live view at 100%. My gear ain't in that generation yet sadly.

[bigdazzler]

AK, i recently tried the d3, 1dsMK4, and the a900 side by side, and the sony did have a better VF. it wasn't something that we were looking for, but it was that apparant that we both commented on it.

yep its like looking through a bay window .. I would find it difficult to go back to a smaller VF now. So much so, that my next major purchase will probably be a second A85O body, for backup, weddings etc ..

sorry reaction ... back on topic now :o

[peterb666]

Slightly OT but is what u see in the VF exactly what the sensor plane sees.
I swear there's been occasions where I've very carefully MF but the point of focus is off on the image. This was done on a large medium format viewfinder too but being film I cannot verify if it's my eye/focussing that's the problem or inaacuracies in camera. We're talking very very shallow dof here though.
I guess the gold standard now is mf in the cameras with live view at 100%. My gear ain't in that generation yet sadly.

The viewfinder does not see what the sensor sees. Instead it it an image that is deflected part the way along the path to the sensor, being bounced off a mirror and through a prism and eyepiece.

The LCD on a live view camera shows exactly what the sensor sees, however this is a processed image an may not be a good representation of the raw image.

[GlennSan]

Also, how do you MF quickly? On a crop body, it takes me 1-2s to find the sharpest point. This makes me useless trying to MF people at an event.:Doh:

I agree with previous posters about pre-focus techniques where possible - pick a point, focus on it and wait for subject to hit that focus mark. Zero lag that way :D

For other types of walk-around images, I use pre-set zone focus where you pick a moderate aperture, use your depth of field scale on the lens to know what your in-focus limits are and shoot away without changing focus at all. You just need to be able to estimate distances well and remember to stay within those distances unless you refocus. Works very well for any wide angle lens and less well as your focal length creeps past standard into telephoto.

Regrettably, many AF lenses have done away with DOF marks on the lens barrel althogether so you're SOL with those with this approach.

I agree with Arthur and also find the small viewfinder on DSLR's hinder accurate fast focussing as does the focus screens fitted as standard to DSLRS. I LOVE using MF on my F4 in particular as it has an exceptional viewfinder. I loath MF on my D100 because it has a small crappy viewfinder and standard focus screen. I also far prefer MF on my original MF lenses rather than using MF mode on my AF lenses. The short amount of rotation from shortest focus distance to infinity on the AF lens helps with AF but I find it makes accurate MF difficult.

That's progress for you ...

[kiwi]

The OP's point is how a lens can be perfect with MF, and off with AF. Usually when this occurs there is a fault and it's not that uncommon, have seen it reported quite often, Can be a calibration issue, or problem with the AF mechanism in general

I often use MF for cricket, radio control cars, macro, etc etc etc

[Kym]

Do any of the DSLRs here have 'catch in focus' feature?
That lets you pre-focus and when the DSLR catches the subject in focus the shutter fires.

[arthurking83]

but to comment on Kiwi's reply.. manual focusing is 'less reliable' than AF on a DSLR, because of the way the screen is set up.

With a proper matte screen fitted to the DSLR body(consumer grade), that is, one with either a focusing assistance area such as a split prism/micro prism collar or a much finer graded matte screen that 'snaps' into focus when a faster than f/5.6 lens is fitted manual focusing is 'more accurate'.

A D90 has live view. If there is any error in the manual or auto focusing systems, use Live View to determine which one is out.
My D300 was a teeny 'micron or two' out in the manual focus system, and there is an adjustment screw in the mirror box that can be adjusted to compensate.

for the OP(if they want to check).

* using a tripod.
* manually focus on a subject to the best point you can with the lens 'wide open'(that is wider than f/5.6) even if you have to use the electronic rangefinder(green confirmation dot) to get focus.
* without touching the focus or moving the camera, switch to live view and zoom to 100% view centering on the subject you focused on.
** Can you focus any better now? If so then your eyesight may be out, the manual focusing system is out(screw adjustment) or an error of judgement in determining DOF(common!).

** do the same thing, but this time using AF.
*Once AF(AF-S mode, with the confirm beep) is determined by the camera/lens, check again with Live View as before. Can you focus better again?

The problem with AF is, as said before, that it depends a lot on which AF point is used, and the amount of contrast in that area, (very importantly!!) the clutter in the chosen AF area box, and the DOF of the subject.
If there are many elements for the AF system to focus on and they're in different DOF zones, and all within the AF area box, then the AF point can be a random event... changing with every AF attempt.

As for which of the different AF type systems are 'more accurate', I'd say that the screw driven AF system would have to more prone to system inaccuracies. Not only do you have the AF system accuracies to battle with, but you have the mechanical connection of the screw driven system to compound the allowable tolerances with the entire AF system of the camera lens.

My (current)screw driven lenses have 'half a millimeter' of play, where you can move the focus collar on the lens and throw the focus point out by a few millimeters(without actually focing the camera's drive motor). ie. mechanical slack.
AF-S systems don't have that, so any focusing errors are entirely due to electrical errors, whether that's the camera/lens communication or a slight error n the SWM motor in the lens. But as there is no mechanical driven connection(aka screw driven connection) there is no slack in the focussing collar.
My Tammy 70-200 has an inbuilt motor in the lens(not SWM type), and it'd have to be the subject ot he same AF inaccuarcies as an AF-S type lens, but with the added complication of it's own internal gearing looseness(slack).. but not the same as the screw driven slack tolerances.
The Tammy(70-200 has no slack in the focusing collar.

Note tho, the old Nikon 80-200/2.8 AF-D was mechanically driven, but had no slack in the focusing collar(as per the other screw driven lenses) but thta was because the focussing collar on the lens locked itself up somehow. That is you couldn't move the focusing collar by hand once the internal gearing was meshed.. whether the lens was mounted to the camera or not!. To release the focusing collar you needed to switch the lens itself to M(manual focus mode) via a dedicated M-A ring on the lens, not as the current AF-S lenses have with their small M/A switches. This lens is either in manual or auto mode, but not both. So to determine how much slack was in the the entire AF system when that lens was mounted was impossible. That lens used to backfocus on my D300(but worked better on the D70s! :confused013).

[ricktas]

Dazz, I don't use AF at all for 95% of my work. But then again, I don't use SFDSLR. When I do use a SFDSLR, then I do autofocus, just because it is too hard to MF with the lenses that i use and those type of cameras. MF is much quicker and more accurate with a rangefinder, so that's what i use for all of my wedding work. i only zone focus in certain situations, and typically when i'm at f1.4 or thereabouts. i think that autofocus cameras get you into the mode of pin point focusing, where in reality, zone focusing is all you ever need. the example i described above for the wedding is more of a trap focus technique.

AK, i recently tried the d3, 1dsMK4, and the a900 side by side, and the sony did have a better VF. it wasn't something that we were looking for, but it was that apparant that we both commented on it.

What does SFDSLR mean?

[bigdazzler]

Small Format Digital Single Lens Reflex I presume ?? :D

[TOM]

:th3:

[Kym]

Small Format Digital Single Lens Reflex I presume ?? :D

4/3? APS-C? APS-H? 35mm?

[bigdazzler]

4/3? APS-C? APS-H? 35mm?

That wasnt the question .. hehe :p

[TOM]

4/3? APS-C? APS-H? 35mm?

correct

[kiwi]

My d3 has trap focus, I believe, not that I've ever tried it

[reaction]

Do any of the DSLRs here have 'catch in focus' feature?
That lets you pre-focus and when the DSLR catches the subject in focus the shutter fires.

The K7 has it, but never tried it. Sounds like it's for catching animals in garden at night type work...

[reaction]

The OP's point is how a lens can be perfect with MF, and off with AF. Usually when this occurs there is a fault and it's not that uncommon, have seen it reported quite often, Can be a calibration issue, or problem with the AF mechanism in general


I have AF issue in 1 lens that uses non SWM/USM motor. The MF on that lens is fine, as is AF/MF on my other lenses, which all use SWM. So I think it’s the lens issue, but not sure.
I’ve heard of Focusing Screens, but don’t get what they do. Why don’t DSLRs come with them installed? I can already see focus in VF, why do I need a Focusing Screen? Sites just show what they sell, but don’t explain how it works…
:efelant:

[arthurking83]

if you look into the mirror box of your camera, and up into the VF prism(towards the flash) you will see an opaque looking screen in there.
This is the focusing matte screen. The opaqueness of it is due to very fine etched lines on it. It obviously decreases the amount of light because it's not totally clear(like a window) but opaque.
This is to create a virtual DOF of the scene that you see through the vf, so that as the part of the scene that you want focused becomes clear in through the vf. If the focusing screen were not there, I suspect that the entire image would be crystal clear, and trying to determine a point of focus would be impossible as the DOF would be unlimited(just as it is looking through a window).

Two things will happen if you don't have a focusing matte screen:
you lose the ability to see the focus as the camera does. The focusing screen is a part of the calibration of the entire optical system, so if you remove it, what you see in focus is not what's actually in focus(dependant on aperture/DOF of course).
The removal of the matte screen creates a tunnel vision effect through the vf. so as you move your eye around there is massive(mechanical?) vignetting to various degrees dependent on how your eye is orientated. I'd say the matte screen acts as a diffuser too.

Remove the matte screen and the image through the vf is instantly brighter and more clear. But there is no perception of DOF at all. Everything appears to be more in focus. You can obviously focus/defocus with the lens, but once the lens is focused to a point, the DOF is much deeper through the vf.

I'm @ home with the kids today, and in between playing Jeeves to them, I'll try to get my focus screen info thread up and running today too.

As for the why you need a focusing screen.. well.. you don't. Well, you don't really need another, because you already got one in your camera(D90 I think?).
These aftermarket focusing screens allow better manual focusing only!
If you opt for the Plus screen from KatZeye, there can or will be exposure issues compared to how each lens's exposure performance was prior to the screen being fitted(that info is available on the Katzeye site too). But once that screen is fitted any exposure compensation required will the same across all lenses.

Which lens is giving you trouble?

[farmer_rob]

Just a brief addition to AK's detailed post - the aftermarket focus screens use a number of optical tricks to make focus "easier" to see. They replace the existing screen in your camera. As AK said, you only need them if you want more accurate manual focusing.

[reaction]

OK yeh, it's for my D90, still don't get why there's magic 'better' screens aftermarket - why not put them in the 1st place?
hmm...

The problem lens is my Tamsy 17-50/2.8 which has a regular motor. It gets close, but I can't MF override :(

[arthurking83]

cost!

I just placed myself in the financially abyss of ordering the Katzeye!

bloody cost over $200 in total.

have a read of my focusing screen tutorial to see some of the why's inherent in standard manufacturer matte screens.

Main reason is cost. But in the uber upper end of the DSLR(and SLR) camera range, the likes of D3's and 1Ds's, the manufacturers have interchangeable matte screens available for those level of cameras.
User changeable and very expensive to boot(even compared to the Katzeye) which is one of the most expensive.

Once my KatzEye is installed, I'll add more info to my Focus screen tute.

EDIT: Oh! the Tammy 17-50!! :confused: This is my only lens that has never caused me any issue whatsoever, at all!
All other lenses have some annoyance or bothersome failing or whatever, except my T17-50. Is it misfocusing with every single shot, or is it misfocusing at a certain focal length?

To determine focusing accuracy you need to start from the start... tripod and shots. Start with test targets of something against a wall.

Could be a lens-camera miscalibration of some kind. Not uncommon. If it's still under warranty take it to get it serviced/checked/calibrated.
if it's the screw driven type, one thing you can do is as I mentioned with my Tammy 28-75mm lens. The lens allows you to twist the focusing collar enough to play with the slack in the system.
Using Liveview lightly twist the focusing collar to see if the focus issue is due to play in the screw drive system. There should be about a millimeter of play in the focus collar from stop to stop.

This is the amount of focus error I see in my Tammy 28-75mm at most portrait distances.

[reaction]

it's misfocusing with almost every single shot, at 2.8. I wouldn't notice it at other f/#
I did some test targets and it's always front focusing. It even looks 'off' in the VF. The VF matches the output focus.
I now try to AF many times before a shot, to hope it gets 'closer' each beep. At 1st I thought it was my fault since it's my 1st f2.8. Anyway, this causes me to lose shots where I move fast and the subject is moving too.
I got it on hols, I don't think Oz will take it under warranty

[para]

when I am photographing wakeboarding from the boat I pre focus the rider no matter what he is doing a 1080 or an ollie(tricks)
He is always at 80ft or whatever he rides at

[arthurking83]

it's misfocusing with almost every single shot, at 2.8. I wouldn't notice it at other f/#....
I got it on hols, I don't think Oz will take it under warranty

makes sense, and .. "bother!"(on getting it on hols)

My favourite test for measuring back/front focus is to do the 'battery test':

Battery test involves 5 batteries(but you can substitute batteries for other similar stuff).

line the batteries up so that the central one is furthest from the lens to the two nest to it, an din turn those two on either side are slightly further back then the two batteries next to them.

So from a birdseye view they're arranged like this:


X
X X
X X
_
camera

(Code is required or I lose the formatting of the X's)

Line up the batteries to be sure that only one battery is seen by any one AF area box.

Take multiple exposures at different apertures up to f/5.6 will be enough.

Take a series of exposures using both vf AF normally, and via Live View. as well.
of course on a tripod. Lighting conditions are irrelevant too. but make sure you wont get any camera vibration like mirror slap, etc.

Begin by using the vf AF system and using the central focus area on the central battery. Determine if the any of the closest batteries on the sides are more in focus than the the central one you focused on.

Then do the same with Live View and zoom to approx 100% view and focus on each battery to see if there is any difference in focus performance(compared to vf AF).

Another thing to check, if you try this test, is to determine if the slack in the screw drive system is part of the issue.
using the vf AF system again, acquire AF, and then lightly twist the focus ring to either side of the slack points to see if there is any difference in focus.

Note tho, if the camera is still set to only shoot once focus is acquired then you may not get any response from the camera as you do these tests.
I use the AF-On button on my D300 to get focus and have separated it from the shutter release. You can set the AE-L button to act as the focus drive to separate it from the shutter release. it''s better to have the focus independent of the shutter release for more user control over the camera.

sorry, I'm rushed because now dinner awaits...... :efelant:

if you had too pay for a service, I think it''d be worth the effort and expense. This lens is wickedly sharp :th3: ... when it works properly of course.

[reaction]

I use a target on screen, cuz it's bright and easy. I take at 45deg left and right, it's easy to see the issue.

On my K-7 it's the same, but not as bad, plus K-7 has focus set, so it's now at -3 and seems fine.

Hoping for a D400 soon...

[reaction]

ok, there is almost NO slack in the motor at all. so it's not that issue.
is there slack in screw types?

[arthurking83]

ok, there is almost NO slack in the motor at all. so it's not that issue....

I would guess that there should be some very small amount of slack in the screw drive system(taking into account both lens and camera).
**BUT I'M NOT 100% SURE!!**(just offer my experience with what I have at my disposal).

Both of my Tammys, 28-75/2.8 and 17-50/2.8 have approximately 1mm or so of slack.
That is, when you rock the focus ring from one end to the other without dragging the camera's screw drive motor, you should feel about 1mm of free play in the focus ring.
At the two end points you need to use much more force to drag the camera's focus motor.
Note you can do this without causing harm.

I only have two screw type lenses to work with, all others are HSM, AF-S, built in lens motor type(Tammy 70-200/2.8) or manual manual focus.

note that some Nikon AF/AF-D lenses which rely on the screw drive may not have any slack in the lenses focus ring as the lenses focus ring locks up(80-200/2.8 AF-D was my one and only lens that did). So, with that lens you can't drag the camera's focus motor even if you wanted too because the lens locked up the focus ring.
With that 80-200mm lens it actually made sense that it operates in that manner too. Sometimes you would end up holding(cradling) the lens from up near the front of it.. right where the focus ring is. So when the lens was focusing the focus ring would not turn. In effect it doubled up as a hand rest, and you wouldn't want to inadvertently move the focus ring once the lens was focused.

[reaction]

well I mailed tammy and they said they'd "change your warranty card to international" by sending it to them. not bad but I still worry about the Sydney side.

[Analog6]

Back before the days of autofocus (yes, I am that old, children!) I did a lot of equestrian photography of moving subjects. Jumping was easy, as someone said, you simply pre-focus. But because the split image screens were more suited to manual focussing, although frustrating with their 'black spots' it was pretty easy to know when you were sharply focussed.

If you wanted to do any extensive manual focussing know you would need to change your screen, it is impossible to do on the modern DSLR ones (IMHO).

[reaction]

Ok so if it makes it a bit dimmer, then in dark indoors it won’t be helpful right?

[arthurking83]

..... it is impossible to do on the modern DSLR ones (IMHO).

Sir!.. surely you jest? :D
(Canon I 'spose? :p)

(with Nikon DSLRs) its the easiest 5 mins spent tampering with stuff.. well, if you like tampering with stuff that is.

Most focusing screens should be an easy change.
With Nikon's you unclip the retaining spring and .... Doh! .. See here, its easier than to repeat it all over again (http://www.ausphotography.net.au/forum/showthread.php?t=61316).

I just got my Katzeye yesterday took some pics of it before fitting it. The tool they send you to release the retaining clip is worth the price alone!(should have thought of that myself!)

vf is massively dark now when I fit the 500 mirror(which is an f/8 lens) and only just workable in semi decent light. The Split prism blackout is indeed as they advertise. Even though this lens is f/8 there is slight darkening of the split prism, but it's nothing as extreme as the dreaded blackspot seen even at f/5.6 in every split prism I've ever used(oh! make that only 3 others :p)
There is zero blackspotting with my one and only f/5.6 lens.

Well worth the money I think(so far).
I'm a bit disappointed that it's darker than I thought it was going to be with the mirror lens.. but there are many and multiple warnings when using a faster focusing screen with slow lenses. Lens is still usable and in a way it gives you a better idea on just how hard this lens is to actually use... because it's such a slow lens
But modern screens are made the way they are so that it's easier to use these slow lenses, which are by far the most common lens type sold in every market around the world... kit lenses, and generally f/5.6 at the long end.

All I need to do now is to undo the manual focus adjustment I made to the camera, so that the split prism is more accurate when manually focusing.
but I have a feeling that the manual focusing system may have been thrown out anyhow as this screen seems to me to be just a tad thicker(material) as the clip was more tightly tensioned when fitting it to the camera.

If you're thinking of getting a (thirdparty)focusing screen then seriously consider the expensive option of the KatzEye, with the OptiBright option.
I'm assuming that without the optibright enhancement a f/5.6 lens would closely resemble my f/8 lens.. which by the way is dark!
With Optibright the f/5.6 lens looks to be close to the standard screen in terms of brightness. I don't have any info on how dark the non optibright screens will look with slow lenses.
They definitely work nice with super fast lenses tho. and there is definitely a difference between the two screens when fitting a super fast f/1.2-f/1.4 lens.
The standard srceen is brighter(not that this actually matters!) but the KatzEye is nicer to focus with(all over the screen, not just the split prism). The image comes into focus much nicer.